HOW THE STAR DID ITS ANALYSIS

The Star has built a complete database of Ontario long-termcare facilities that compares the Ministry of Long-Term Care’s public data on active and resolved COVID-19 outbreaks to records sourced from the Long-Term Care public reporting database, the ministry’s home-by-home data on bed licensing standards and the Star’s own data on local population-adjusted infection rates.

The Star first used this database to create a home-byhome record of every reported COVID-19 outbreak, from which we calculated each home’s per capita death rate, defined as the number of reported COVID-19 deaths for every 100 registered beds.

The Star then classified the homes by six overlapping characteristics, which we chose based on the industry’s claims, the existing research and the available data.

The categories the Star examined are:

Ownership type, defined as for-profit, non-profit and municipal, as listed in the public reporting database.

Design age, in which an older home is any facility where most beds meet or fall below 1972 design standard. (This is the same definition used in the Canadian Medical Association Journal paper from last summer and in the recent Ministry of Health analysis.)

Local infection rate, based on the local health unit’s cumulative COVID-19 cases, adjusted for population. For the full pandemic, we defined any health unit with more than 2,000 cumulative cases per 100,000 people as having a high infection rate; a unit with between 1,000 and 2,000 is medium; less than 1,000 cases per 100,000 is low.

Chain ownership, based on the home’s listed licensee. A home with a unique licensee is considered a single home, a licensee with two to nine facilities is a chain, and a licensee with 10 or more homes is considered a large chain.

Facility size, in which less than 100 beds is small, between 100 and 199 is medium, and more than 200 is large.

Management firm, based on the existence of a listed outside manager in the LTC public reporting database.

The Star next conducted a “stratified analysis” comparing the average death rates in homes sorted by all of these factors isolated and in combinations of up to three — a total of 81 separate comparisons.

For-profit facilities reported higher death rates than the average for non-profit and municipal homes in all but three comparisons, one of which saw no deaths at all.

The Star’s key finding is that for-profit facilities reported the most deaths even comparing design age and local infection rate in combination. For-profits reported the highest death rates in older facilities, whether in high-, medium- or low-incidence communities, and also in newer homes, whether in high-, medium- or low-incidence communities.

The result shows that even looking only at homes by the factors the for-profit sector has consistently said create the highest risk for worse outcomes, for-profits nonetheless performed much worse on average.

(The Star’s analysis otherwise agrees with the finding that facility age and local context have both been important factors in death rates.)

After presenting our findings to researchers, the Star reran the analysis several different ways prior to publication, finding consistent results.

Most notably, we considered the claim that the unique circumstances of some of Ontario’s worst-hit homes could be skewing the trend. To adjust for this possibility, we ran the analysis again, excluding Ontario’s 10 worst-hit homes by deaths on a per-bed basis.

Even though these 10 are themselves disproportionately for-profit, the key result was consistent: for-profits reported the worst death rate, regardless of design age or local infection rate.